The Biggest Modeling Problem We Don’t Talk About

Financial modeling is both art and science, and while most people focus on the complexities of building models, I want to give some air time to an equality important issue: the aesthetics of presenting a model so that someone can understand how the model actually works.

Human-readability is a huge issue in modeling. Just check out this screenshot:

Imagine you are opening this 25MB file for the first time. You are going to hit a number of roadblocks:

1) Read-only: Given the size on the file, hitting the “Edit Anyway” or “Enable Content” buttons are a recipe for disaster. There is a decent chance your Excel program will crash and, if it doesn’t, calculations can take up to a minute to run given the volume of data and layering of formulas with which Excel is trying to contend.

2) Repeated cell values: The biggest cause of the massive file size is the repeated data stored all over the file (such as dates duplicated on every tab). A spreadsheet model stores whatever content you type into each cell, so a ton of data and compute power is spent on duplicative scaffolding and limit the performance of Excel’s most powerful features.

3) Human-readability: Spreadsheet models are inherently difficult for third-parties to understand. See the formula in cell DX3886? Good luck trying to figure out what this means:

=+IFERROR(IF($F3886="Recurring",$Q3886*DX3875,IF($F3886="Time Rate",($Q3886/4.33)*DX$5/5,$Q3886))*PRODUCT(1+IF($R$10:$V$10<=IF(INDEX(DX$6>=Client_SiteMonth,MATCH($D3886,Client_Names,0),MATCH(DX$8,Client_Months,0)),DX$8,DX$8-1),$R3886:$V3886)),0)

3a) Did you finish step 3? Congratulations! Your prize is the ability to check the other 242,089 cells in this tab

3b) There are 36 tabs (some of which might be hidden)

Does this inspire confidence?

Everyone makes mistakes. And whether you are first-time founder, CFO with 40 years’ experience, seasoned investor, or a 2nd year investment banker, Excel’s lack of emphasis on human readability causes errors to go uncaught and overlooked. Expert modelers use a number of conventions to try to improve human readability, from color coding cells based on their formula type (blue = hardcoded assumption, green = formula that references cells on other tabs), yellow highlighting, etc. Nevertheless, the beauty and danger of Excel is that literally anything can be anywhere. And as a result even the most well thought-out and formatted models requires hours if not days to fully understand. And even then, reviewers need need call and emails to clarify what, exactly, is going on in certain places.

Excel’s lack of emphasis on human readability causes errors to go uncaught and overlooked.

We believe in a future where the assumed tradeoff between the flexibility of Excel and the human-readability of your model isn’t a tautology. Finance professionals should demand more from the interfaces they use without sacrificing the flexibility that Excel grants them. If you’re reading this and share our frustrating with Excel’s lack of human readability, we’d be excited to connect.

Prophet’s interface was built with human readability as a core tenet. We also understand that this means different things to different types of users. For the “architects” who build and own models, this means smart templates to reduce duplicative tasks like building working capital schedules. For the “collaborators” who need to input certain assumptions but don’t want to be overwhelmed by an arcane Excel template, this means an intuitive input interface. And most importantly, for “reviewers” like management, board members and investors this means formulas that are written to natively reference things, not cells. Revenue = Recurring Revenue + Consulting Revenue is a lot more helpful than ‘Operating Model’!$E$12 = ‘Operating Model’!$E$10 + SUMIF(‘Financial Model’!$M$12:$M$118,”Consulting Revenue”,‘Financial Model’!$Q$12:$Q$118).

In contrast to Excel, Prophet was built to empower architects, collaborators and reviewers alike to improve all facets of financial modeling. Our smart templates reduce the time it takes to build v1. The friendly input modules and system integrations reduce iterative data collection, and the human readable interface reduces the time to review, approve and analyze the outputs.